We have seen in the sections above that prior knowledge will not support new learning if it is insuffi cient or inappropriate for the task at hand. But what if it is downright wrong? Research indicates that inaccurate prior knowledge (in other words, fl awed ideas, beliefs, models, or theories) can distort new knowledge by predisposing students to ignore, discount, or resist evidence that confl icts with what they believe to be true (Dunbar, Fugelsang, & Stein, 2007 ; Chinn & Malhotra, 2002 ; Brewer & Lambert, 2000 ; Fiske & Taylor, 1991 ; Alvermann, Smith, & Readance, 1985 ). Some psychologists explain this distortion as a result of our striving for internal consistency. For example, Vosniadou and Brewer (1987) found that children reconcile their perception that the earth is fl at with formal instruction stating that the earth is round by conceiving of the earth as a pancake: circular but with a fl at surface. In other words, children — like all learners — try to make sense of what they are learning by fi tting it into what they already know or believe.
Inaccurate prior knowledge can be corrected fairly easily if it consists of relatively isolated ideas or beliefs that are not embedded in larger conceptual models (for example, the belief that Pluto is a planet or that the heart oxygenates blood). Research indicates that these sorts of beliefs respond to refutation; in other words, students will generally revise them when they are explicitly confronted with contradictory explanations and evidence (Broughton, Sinatra, & Reynolds, 2007 ; Guzetti, Snyder, Glass, & Gamas, 1993 ; Chi, 2008 ). Even more integrated — yet nonetheless fl awed — conceptual models may respond to refutation over time if the individual inaccuracies they contain are refuted systematically (Chi & Roscoe, 2002 ).
However, some kinds of inaccurate prior knowledge — called misconceptions — are remarkably resistant to correction. Misconceptions are models or theories that are deeply embedded in students ’ thinking. Many examples have been documented in the literature, including na ï ve theories in physics (such as the notion that objects of different masses fall at different rates), “ folk psychology ” myths (for example, that blind people have more sensitive hearing than sighted people or that a good hypnotist can command total obedience), and stereotypes about groupsof people (Brown, 1983 ; Kaiser, McCloskey, & Proffi tt, 1986 ; McCloskey, 1983 ; Taylor & Kowalski, 2004 ).
Misconceptions are diffi cult to refute for a number of reasons. First, many of them have been reinforced over time and across multiple contexts. Moreover, because they often include accurate — as well as inaccurate — elements, students may not recognize their fl aws. Finally, in many cases, misconceptions may allow for successful explanations and predictions in a number of everyday circumstances. For example, although stereotypes are dangerous oversimplifi cations, they are diffi cult to change in part because they fi t aspects of our perceived reality and serve an adaptive human need to generalize and categorize (Allport, 1954 ; Brewer, 1988 ; Fiske & Taylor, 1991 ).
Research has shown that deeply held misconceptions often persist despite direct instructional interventions (Ram, Nersessian, & Keil, 1997 ; Gardner & Dalsing, 1986 ; Gutman, 1979 ; Confrey, 1990 ). For example, Stein and Dunbar conducted a study (described in Dunbar, Fugelsang, & Stein, 2007 ) in which they asked college students to write about why the seasons changed, and then assessed their relevant knowledge via a multiple choice test. After fi nding that 94 percent of the students in their study had misconceptions (including the belief that the shape of the earth ’ s orbit was responsible for the seasons), the researchers showed students a video that clearly explained that the tilt of the earth ’ s axis, not the shape of the earth ’ s orbit, was responsible for seasonal change. Yet in spite of the video, when students were asked to revise their essays, their explanations for the seasons did not change fundamentally. Similarly, McCloskey, Caramazza, and Green (1980) found that other deeply held misconceptions about the physical world persist even when they are refuted through formal instruction.
Results like these are sobering. Yet the picture is not altogether gloomy. To begin with, it is important to recognize that conceptual change often occurs gradually and may not be immediately visible. Thus, students may be moving in the direction of more accurate knowledge even when it is not yet apparent in their performance (Alibali, 1999 ; Chi & Roscoe, 2002 ). Moreover, even when students retain inaccurate beliefs, they can learn to inhibit and override those beliefs and draw on accurate knowledge instead. Research indicates, for instance, that when people are suffi ciently motivated to do so, they can consciously suppress stereotypical judgments and learn to rely on rational analysis more and stereotypes less (Monteith & Mark, 2005 ; Monteith, Sherman, & Devine, 1998 ). Moreover, since consciously overcoming misconceptions requires more cognitive energy than simply falling back on intuitive, familiar modes of thinking, there is research to suggest that when distractions and time pressures are minimized, students will be more likely to think rationally and avoid applying misconceptions and fl awed assumptions (Finucane et al., 2000 ; Kahnemann & Frederick, 2002 ).
In addition, carefully designed instruction can help wean students from misconceptions through a process called bridging (Brown, 1992 ; Brown & Clement, 1989 ; Clement, 1993 ). For example, Clement observed that students often had trouble believing that a table exerts force on a book placed on its surface. To help students grasp this somewhat counterintuitive concept, he designed an instructional intervention for high school physics students that started from students ’ accurate prior knowledge. Because students did believe that a compressed spring exerted force, the researchers were able to analogize from the spring to foam, then to pliable wood, and fi nally to a solid table. The intermediate objects served to bridge the difference between a spring and the table and enabled the students to extend their accurate prior knowledge to new contexts. Using this approach, Clement obtained signifi cantly greater pre - to posttest gains compared to traditional classroom instruction. In a similar vein, Minstrell ’ s research (1989) shows that students can be guided away from misconceptions through a process of reasoning that helps them build on the accurate facets of their knowledge as they gradually revise the inaccurate facets.
Implications of This Research It is important for instructors to address inaccurate prior knowledge that might otherwise distort or impede learning. In some cases, inaccuracies can be corrected simply by exposing students to accurate information and evidence that confl icts with fl awed beliefs and models. However, it is important for instructors to recognize that a single correction or refutation is unlikely to be enough to help students revise deeply held misconceptions. Instead, guiding students through a process of conceptual change is likely to take time, patience, and creativity.
